May 31, 2026 – Ann Arbor, MI – Non-Contact Sleep Monitoring Using PIR
Researchers at the University of Michigan have published a study demonstrating that ceiling-mounted PIR sensors can track sleep patterns without requiring wearables or cameras.
The system uses a single high-sensitivity PIR sensor mounted above the bed. Machine learning algorithms analyze micro-movements during sleep to determine sleep stages, restlessness, and breathing patterns.
In a study with 50 participants, the system achieved:
- Sleep stage classification accuracy: 85% (compared to EEG as gold standard)
- Sleep duration accuracy: ±10 minutes
- Awake detection sensitivity: 90%
“This could enable long-term sleep monitoring in homes without any devices attached to the body,” said lead researcher Dr. Sarah Chen. “The privacy advantages over cameras are significant.”
The research was published in the Journal of Medical Internet Research. The team is seeking industry partners for commercialization, with products estimated to be 2-3 years away.